Journal of Animal Science, 2023, 101, 1–11

https://doi.org/10.1093/jas/skac422
Advance access publication 28 December 2022
Non ruminant nutrition

The effect of harvest time of forage on carbohydrate

digestion in horses quantified by in vitro and mobile bag
techniques

Downloaded from https://academic.oup.com/jas/article/doi/10.1093/jas/skac422/6964640 by guest on 08 December 2023

Frida Lindskov Stang,†,2 Rikke Bjerregaard,†,2 Cecilia Elisabeth Müller,‡ Åshild Ergon,‖
Magnus Halling,$ Nana Wentzel Thorringer,¶ Alemayehu Kidane,¶, and
Rasmus Bovbjerg Jensen¶,1
†
Department of Veterinary and Animal Sciences, University of Copenhagen, DK-1870 Frederiksberg, Denmark
‡
Department of Animal Nutrition and Management, Swedish University of Agricultural Sciences, SE-750 07 Uppsala, Sweden
‖
Department of Plant Sciences, Norwegian University of Life Sciences, NO-1430 Ås, Norway
$
Department of Crop Production Ecology, Swedish University of Agricultural Sciences, SE-750 07 Uppsala, Sweden
¶
Department of Animal and Aquacultural Sciences, Norwegian University of Life Sciences, NO-1430 Ås, Norway
Both authors contributed equally.
2

Corresponding author: rasmus.bovbjerg.jensen@nmbu.no

1

Abstract
Carbohydrates in forages constitute an important part of the feed ration for all horses. The aim of the present study was to investigate the effect
of harvest time on carbohydrate composition and digestion of various grass species. The experiment was divided into three parts 1) charac-
terization of the chemical composition of experimental feeds (6 grass species: meadow fescue [MF], cocksfoot [CF], perennial ryegrass [PR],
smooth bromegrass [SB], tall fescue [TF], and timothy [TI], and 3 harvest times: early, medium, and late first cut), 2) measurements of the in
vitro digestion of selected experimental feeds (the 6 grass species, and 2 harvest times [early and late]) measured by in vitro gas production,
and 3) in vivo digestion of selected experimental feeds (2 grass species: CF and PR, 2 harvest times [early and late]) measured by the mobile
bag technique using caecum cannulated horses. An experimental field was established with plots containing each of the grass species in three
replicate blocks. Grass samples were cut between 1200 and 1400 h at 4th of June (early first cut), 17th of June (medium first cut), and 1st of
July (late first cut) and analyzed for crude protein (CP), neutral detergent fiber with heat stable amylase and free of residual ash (aNDFom) and
water-soluble carbohydrates (WSC). The in vitro fermentation was investigated using the ANKOM RF gas production technique, where feeds
were incubated for 48 h using horse caecal fluid as an inoculum. Gas production was modeled, and maximum gas production (MGP) was used
to evaluate the potential digestibility of the feeds. Based on the chemical analyses and the in vitro experiment, early and late harvested CF
and PR were selected for the in vivo experiment, which was conducted as a randomized 4 × 4 Latin square design including four periods, four
horses and four feeds. In general, the CP content decreased whereas the aNDFom content increased as the grasses matured. The content
of WSC increased in SB and TI, but decreased in CF, and fructans increased in SB, TI, PR, and TF as they matured. The in vitro MGP showed a
clearer difference between harvest times than between grass species. Harvest time had larger effect on digestibility than grass species, and a
high precaecal disappearance of the WSC fraction was measured by the mobile bag technique. Cocksfoot was identified as a grass species with
potentially low digestibility and low WSC content and could potentially be used more for horses.

Lay Summary
Feedstuffs contain different carbohydrate fractions that are digested in different parts of the gastrointestinal tract of horses. Grass for grazing
or harvesting contains variable amounts of structural carbohydrates such as cellulose and hemi-cellulose (named fibres) and nonstructural car-
bohydrates which in temperate grass species include sugars and fructans (named water soluble carbohydrates (WSC)). This study quantified
carbohydrate composition and digestion of six grass species (perennial ryegrass, timothy, smooth bromegrass, tall fescue, cocksfoot, and
meadow fescue) harvested at three different times (early, medium, and late) and preserved as hay. In general, fiber content increased as the
grasses matured, whereas WSC content varied to a large extent. In vitro fermentation using horse caecal fluid was used to quantify digestion
of early and late cut grass samples of all species. Harvest time (early vs. late) had a larger effect on in vitro fermentation compared to the effect
of grass species. Early and late harvested perennial ryegrass and cocksfoot were further selected for detailed studies of precaecal digestion in
vivo as these species had highest and lowest WSC content. In general, cocksfoot was identified as grass species with low digestibility and low
WSC concentration compared to the other species investigated.
Key words: equine, fructans, grass species, in situ, sugar, water-soluble carbohydrates
Abbreviations: A, asymptotic gas production (mL/g DM); B, time after incubation at which half of the maximum gas production is reached; C, constant related
to the shape of the curve; CF, cocksfoot; CP, crude protein; G, gas produced (mL/g DM); MGP, maximum gas production; MF, meadow fescue; Mi, bags found at
the ith time as a fraction of the total number of bags found; NDF, neutral detergent fibre; aNDFom, NDF assayed with heat-stable amylase and expressed without
residual ash; NSC, Non-structural carbohydrates; PR, perennial ryegrass; SB, smooth bromegrass; TF, tall fescue; t, time (h); tRM, time elapsed from intubation to
the midpoint of the ith time; TI, timothy; tRM, time (h) for maximum rate of digestion; WSC, water-soluble carbohydrates

Received November 8, 2022 Accepted December 21, 2022.

© The Author(s) 2022. Published by Oxford University Press on behalf of the American Society of Animal Science.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/),
which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.
2 Journal of Animal Science, 2023, Vol. 101

Introduction the grass matures (Ragnarsson and Lindberg, 2008, 2010;

Müller, 2012). Horses with high energy requirements there-
Carbohydrates in forage grasses are divided into structural
fore benefit from highly digestible early harvested forage
(mainly cellulose, hemicelluloses, and pectin) and non-
(Ringmark et al., 2013), whereas horses with comparably
structural (sugars and starch). In cool-season C3-grasses,
low energy requirements benefit from a late harvested for-
the non-structural carbohydrates (NSC) consist mainly of
age with a low digestibility (Harris et al., 2017). However,
glucose, fructose, sucrose and fructans, together termed
the grass species can affect both fiber content and digest-
water-soluble carbohydrates (WSC). Grasses produce simple
ibility despite being harvested in the same plant maturity
sugars from photosynthesis, and when produced in excess of
(Särkijärvi et al., 2012). This knowledge could be used to
the energy requirements of the plant for growth and devel-
a higher degree in equine nutrition, but more information
opment, sugars are converted to storage carbohydrates in the
about differences in fiber digestibility between grass species
plant (Longland and Byrd, 2006). Fructan is the primary stor-
is needed, especially for harvest times that are relevant for

Downloaded from https://academic.oup.com/jas/article/doi/10.1093/jas/skac422/6964640 by guest on 08 December 2023

age carbohydrate of cool-season C3-grasses. Fructans consist
horse requirements.
of a terminal glucose with additional fructose units linked
Various in vitro and in vivo techniques can be used to
together in branched or non-branched structures of various
quantify digestion of forages in horses, each with different
lengths; different grass species having different types (Pollock
advantages and disadvantages. One technique commonly
and Cairns, 1991).
used is the in vitro gas production technique (IVGPT) where
The carbohydrates constitute a large proportion of the for-
a positive correlation between the produced gas and digest-
age and are important energy sources for all horses. The struc-
ibility has been found when using fecal contents or digesta
tural carbohydrates are fermented by hindgut microbes to
as inoculum (Lowman et al., 1999). A disadvantage of the
short-chained fatty acids (SCFA) providing the largest energy
IVGPT is that it cannot quantify where in the gastrointesti-
source for the horse (Janis, 1976), and the WSC are mainly
nal tract of the horse the different carbohydrate fractions are
digested and absorbed as glucose and fructose in the small
digested. As this is of interest particularly for fructans, other
intestine (Dyer et al., 2002; Merediz et al., 2004). Currently
techniques are required. The mobile bag technique (MBT) in
it is not clearly known how and where fructans are digested
caecum-cannulated horses can be used to quantify the precae-
within the equine gastro-intestinal tract. It is assumed that
cal digestibility of feedstuffs and its components, even if most
fructans cannot be digested in the small intestine because
research using this technique has focused on starch digestion
horses lack the enzymes that break the linkages between the
of grains (Philippeau et al., 2014; Thorringer et al., 2020)
fructose molecules in fructans. Nevertheless, preliminary
or fiber digestion (Moore-Colyer et al., 2002; Thorringer et
studies have shown that part of fructans might be digested
al., 2022). It is possible to use the MBT also for digestibility
in the stomach of the horse (Coenen et al., 2006), probably
studies of WSC, but as of today it is not known what the
through acid hydrolysis from the acidic environment in the
optimal pore size of mobile bags should be and how it affects
lower part of the stomach. Fructans have also been linked
the results.
to development of laminitis in horses (Longland and Byrd,
The aim of the present study was therefore to quantify
2006), and it has also been reported that high doses (10 g/kg
carbohydrate composition of various grass species har-
body weight) of nongrass fructans resulted in hindgut acido-
vested at different times and their digestion in horses using
sis and laminitis in horses (Van Eps and Pollitt, 2006). Hence,
IVGPT and MBT, the latter with different pore sizes of
more research is needed to understand the in vivo digestion
bags. The hypotheses were 1) carbohydrate composition is
of grass fructans in horses. Sugars causing a glucose and insu-
affected by grass species and harvest time (fiber content
lin response are also of interest in equine nutrition, especially
increases and WSC decreases as the plants mature), 2) in
for horses with insulin dysregulation (Frank et al., 2010;
vitro digestion is affected more by stage of maturity than
Lindåse et al., 2016), pituitary pars intermedia dysfunction
grass species, 3) the in vivo digestion of glucose, fructose
(McGowan, 2008) or polysaccharide storage myopathy
and sucrose but not fructans is complete in the small intes-
(Firshman et al., 2003; Borgia et al., 2011). These horses’ risk
tine of the horse, and 4) a mobile bag pore size of 36 µm
severe health problems with high intake of NSC and should
will results in higher loss of nutrients than a pore size of
therefore ideally be fed a diet containing less than 10% NSC
15 µm.
of dry matter (DM), preferably lower (Borgia et al., 2011;
Lindåse et al., 2018). Recommendations on how to produce
forage with NSC (or WSC) concentrations below 10% of DM Materials and Methods
are currently lacking, as most of the research performed on
forages and their composition and nutritive values have been Experimental design
focusing on ruminants where a high NSC content is desired All housing, management, and experimental procedures fol-
(Humphreys et al., 2006; Shewmaker et al., 2006). Different lowed the laws and regulations for experimental animals in
grass species may differ in their capability to produce forages Norway (i.e., Regulations on the Use of Animals in Experi-
with less than 10% NSC of DM (King et al., 2012; Jensen et ments, July 2015), and the experiment was approved by the
al., 2014) and more information is needed on how different Norwegian Food Safety Authority (FOTS ID 22251). The
grass species respond to different plant maturity at harvest in experiment was divided into three parts 1) characterization
their NSC concentration and composition. of the chemical composition of experimental feeds (6 grass
The structural carbohydrates and fiber digestibility species, 3 plant maturities), 2) in vitro digestion of selected
should be considered when selecting grass species for for- experimental feeds (6 grass species, 2 plant maturities) mea-
ages for different types of horses. As fiber content increases sured by the IVGPT, and 3) in vivo digestion of selected exper-
digestibility decreases, and it has a major impact on the imental feeds (2 grass species, 2 plant maturities) measured by
digestible energy value of the forage which decreases as the MBT using caecum cannulated horses.
Lindskov Stang et al.
Experimental feeds
The grass species used for the experiments were grown
at the Norwegian University of Life Sciences, Ås, Nor-
way (59°39ʹ42.5″N 10°44ʹ56.0″E, 82 m above sea level,
weather data is presented in Supplementary Figure S1). The
species sown were meadow fescue (MF, Festuca pratensis,
cv. Minto), cocksfoot (CF, Dactylis glomerata, cv. Laban),
perennial ryegrass (PR, Lolium perenne, cv. Figgjo), smooth
bromegrass (SB, Bromus inermis, cv. Leif), tall fescue (TF,
Festuca arundinacea, cv. Swaj), and timothy (TI, Phleum
pratense, cv. Grindstad). An experimental field was estab-
lished with a total of 36 plots divided into three blocks,
each block containing two randomly placed replicates of
each grass species. The grasses were sown at the end of May
2018 with barley (Hordeum vulgare, cv. Salome) as a cover
crop at a seed rate of 180 kg/ha and irrigated during the
summer 2018. The plots were fertilized with a compound
fertilizer at a rate of 80, 30, and 100 kg N ha−1 prior to
sowing, after harvesting the cover crop, and in spring 2019,
respectively. The grasses were cut with a Haldrup plot har-
vester at approximately 7 cm height in autumn 2018, and
in 2019 the grasses were cut between 1200 and 1400 h at
4th of June (early first cut), 17th of June (medium first cut),
and 1st of July (late first cut). One of the two replicate plots
within a block was used for the early cut, while the other
replicate plot was divided in two: one half for the middle
cut and one half for the late cut. This resulted in three rep-
licate samples of each grass species for each of the three
harvest times. The biomass from each plot was mixed, and
approximately 1 kg of the sample was oven-dried at 50 °C
for 3 d, then milled to pass a 1-mm screen (cutting mill SM
200, Retch GmbH, Haan, Germany) and stored for later
analysis.
Animals and diets
Five caecum cannulated, 14- to 26-yr-old Norwegian Cold-
blood trotter geldings with an average initial BW of 575 kg
(range 545 to 631 kg) were used for the in vitro and in vivo
experiments. The horses were fitted with a permanent polyvi-
nyl chloride cannula (length ~ 15 cm; 40 mm o.d., and 30 mm
i.d.) at the base of the caecum close to the ileo-caecal junc-
tion more than 10 yr before the experiment. The horses were
housed in an unheated barn in 3 × 3 m individual box stalls
with rubber mats and wood shavings as bedding-material.
They were allowed access to a gravel paddock for approx-
imately 6 h per day, except for days with in vivo measure-
ments where the horses were allowed access to the paddock
for approximately two hours after the measurements. The
horses received 9 to 12 kg hay as-fed (timothy-meadow fes-
cue) divided into three equal meals fed at 0600, 1400, and
1930 h (the morning meal was fed between 0700 and 0745 h
on test days in the in vivo experiment). The morning meal
also included 30 g/d sodium chloride and 100 g/d of a com-
mercial mineral and vitamin supplement (Champion Mul-
titilskudd, Felleskjøpet Fôrutvikling, Trondheim, Norway)
consisting, per kg original matter, of Ca, 100 g; P, 70 g; Mg,
32 g; Cu, 840 mg; NaCl, 50 g; Zn, 2830 mg; Mn, 1530 mg;
Fe, 2460 mg; I, 18 mg; Co, 6 mg; Se, 10.2 mg vitamin A,
107,000 I.U.; vitamin D, 11,300 I.U.; vitamin E, 9,600 mg;
vitamin B1, 260 mg; vitamin B2, 120 mg; vitamin B6, 100
mg: vitamin B12, 0,8 mg; niacin, 270 mg; folic acid, 150 mg;
biotine, 15 mg, and vitamin C, 270 mg. Water was available
3
ad libitum from automatic water troughs in the individual
box stalls, and from buckets in the gravel paddock.
In vitro experiment
One of the three replicates of the early and late harvested grass
samples (MF, CF, PR, SB, TF, and TI) were randomly selected
for the in vitro experiment. The ANKOM RF gas production
system (Version 9.8.3, ANKOM Technology, Macedon, NY,
USA) was used to measure the in vitro digestibility accord-
ing to the recommendations from the company (ANKOM,
2022). In brief, three replicates of each of the 12 grass sam-
ples (6 grass species and 2 plant maturities), three replicates
Downloaded from https://academic.oup.com/jas/article/doi/10.1093/jas/skac422/6964640 by guest on 08 December 2023
without feedstuffs (blanks), and three replicates with an inter-
nal standard (sugar beet pulp) were incubated using caecal
inoculum (42 bottles in total). Approximately 1.1 g of each
feedstuff was weighed into 250 mL incubation bottles. A final
buffer solution was prepared by mixing 875 mL macro min-
eral-, 875 mL buffer-, 0.44 mL micro mineral-, and 4.4 mL
resazurin-solution in addition to 1,750 mL distilled water
according to Goering and Van Soest (1970) omitting the addi-
tion of trypticase. The final buffer solution was flushed with
CO2, kept in a heated water bath at 39 °C, and 183 mL reduc-
ing solution was added to the final buffer solution. The final
buffer solution was flushed with CO2 and a reducing solution
was added, and the color changed from purple to clear, indi-
cating that the solution was saturated with CO2. Each bottle
with feed samples and the three blanks were filled with 66 mL
of the final buffer solution and placed in a heating oven at 38
°C while caecal fluid inoculum was prepared.
Caecal inoculum was collected from three horses through
their caecal cannula approximately five hours after the morn-
ing meal and stored in preheated thermos bottles for the
two-minute transportation between the stable and the labo-
ratory. The inoculum was strained through a precision woven
synthetic monofilament fabric with a pore size of 200 μm
(Nitex 03-200/47, SEFAR, Heiden, Switzerland) into a bottle
placed in a heated water bath at 39 °C. Then 34 mL inoculum
was filled into each bottle containing feed samples and final
buffer solution, the headspace of the bottles was flushed with
CO2 and then closed with the head module. The bottles were
placed in an incubator (B 8420, TERMAKS, Bergen, Norway)
at 39 °C on two separate shakers (Gyro rocker SSL3, STU-
ART, Staffordshire, United Kingdom) rotating at 11 rounds
per minute, and measurements started. The ANKOM RF gas
production system was set to record every 10 minutes and to
release gas when the pressure in the bottles exceeded 0.75 psi.
After 48 h the incubation was terminated, and the pH of each
bottle was measured (pH 3310, Xylem Analytics, Weilheim,
Germany). The content of each bottle was poured into pre-
weighed polyester bags (Saatifil PES 12/6, Saati S.p.A., Appi-
ano Gentile, Italy) with a pore size of 12 μm and a surface
area of approximately 200 cm2. During this process, the filled
bags were kept in water at 5–10 °C, after which all bags were
washed for 35 min using the cold-water wool-programme
with no final spin in a domestic washing machine (Avantixx
7 VarioPerfect, BOSCH, Gerlingen Schillerhöhe, Germany).
Bags were dried in a heating oven (TS 8430, TERMAKS, Ber-
gen, Norway) at 45 °C for 48 h after which each bag was
weighed.
Mobile bag experiment
The in vivo experiment was conducted as a random-
ized 4 × 4 Latin square design over a four-week period
4 Journal of Animal Science, 2023, Vol. 101
i­ncluding four horses and four feed samples. The feed sam-
ples included were the early and late harvests of PR and
CF. In addition to the different feed samples, two different
pore sizes of the mobile bags were used (15 and 36 μm). In
total, 352 mobile bags were used, with 40 replicates of each
grass × harvest × pore size combination, and 4 replicates of
each combination, which were used to determine the wash-
ing loss of nutrients from the bags. The size of the mobile
bags was 1 × 2 × 12 cm and they were prepared as described
by Thorringer et al. (2020) from a precision woven synthetic
monofilament fabric with a pore size of either 15 μm (Nitex
03-15/10, SEFAR, Heiden, Switzerland) or 36 μm (Nitex
03-36/28, SEFAR, Heiden, Switzerland). A steel washer with
a diameter of 10 mm was heat sealed into one end of each
bag, except for bags used to determine washing loss. Each
mobile bag was weighed, marked for identification, and
filled with 0.5 g of feed, ensuring a feed to surface area (FSA)
of 21 mg/cm2.
The mobile bags were intubated each week between 0700
and 0745 h before the morning meal. The bags (20 bags
per horse per period) were soaked in cold tap water before
they were administered into the stomach with a nasogastric
tube flushed with approximately 2 L of tap water. To cap-
ture the mobile bags, a 50 cm long nylon tube containing
a double-sided magnet with a diameter of 2 cm was intro-
duced into the caecum through the cannula. The magnet was
withdrawn every hour until 10 h post intubation, and bags
not recovered at this time were collected in feces two to four
times per day until all bags had been recovered. After collec-
tion bags were quickly rinsed with tap water and then stored
at −20 °C. The mobile bags were thawed at room tempera-
ture, placed in a washing bag (35 × 28 cm, pore size approx-
imately 1 mm), and washed for 35 min using the cold-water
wool-programme with no final spin in the same domestic
washing machine as described previously. Each washing bag
contained mobile bags collected from the caecum of one
horse (12–20 bags), and two washing bags were washed at
the same time. Washing loss from mobile bags that had not
been intubated was done in a similar way. Bags were then
dried in a heating oven (TS 8430, TERMAKS, Bergen, Nor-
way) at 45 °C for 48 h after which each bag was weighed. To
ensure enough feed residue for chemical analysis, the mobile
bags were pooled by each grass species × harvest time × pore
size combination in addition to the three time-intervals
where bags were found in the caecum (1–3, 4–6, and 7–10 h
after bags were administered into the stomach) before fur-
ther chemical analysis.
Chemical analysis
The experimental feeds and feed residues from the in vitro
and in vivo experiments were analyzed for dry matter (DM)
content by drying to a constant weight for 24 h at 103 °C
(feeds) or for 48 h at 45 °C (feeds and feed residues), and
samples were incinerated at 550 °C for 16 h for ash determi-
nation (feeds). The content of NDF was determined using an
ANKOM200 Fibre Analyzer (ANKOM Technology, Mace-
don, New York, USA) using heat-stable amylase followed by
combustion at 550 °C and expressed without residual ash
(aNDFom). Nitrogen was determined using a Kjeltec TM
8400 (FOSS, Hillerød, Demark) and crude protein (CP) con-
tent was calculated as N × 6.25. Analysis of WSC, glucose,
fructose, sucrose, and fructans were performed using an enzy-
matic-spectrophotometric method described by Udén (2006).
Calculations and statistical analysis
Calculations for the in vitro experiment
The cumulative gas production measurements were converted
from psi to moles by the ideal gas law:
V
n=p×
R·T
where n is the gas produced (moles), p is the pressure (kPa),
V is the head-space volume in the bottles (L), R is the gas
constant of 8.314472 (L ∙ kPa ∙ K−1 ∙ mol−1), and T is the tem-
perature (°K). Gas produced was then converted from moles
Downloaded from https://academic.oup.com/jas/article/doi/10.1093/jas/skac422/6964640 by guest on 08 December 2023
to mL by the following formula:
Gas produced (mL) = n × 22, 400 mL/mole
where n is the gas produced (moles) and 1 mole of gas will
occupy 22,400 mL at standard conditions (273.15 °K and
101.325 kPa). The gas produced in mL was then corrected
for the DM content in the bottle and fitted to the monophasic
model by Groot et al. (1996) using the NLIN procedure in SAS
(version 9.4, SAS Institute Inc., Cary, North Carolina, USA):
n
 Ai
G= C
Bi i
i=1 1+ tCi
where G is the gas produced (mL/g DM), A is the asymptotic
gas production (mL/g DM) referred to as maximum gas pro-
duction (MGP), B is the time (h) after incubation at which half
of the MGP is reached, C is a constant related to the shape of
the curve, and t is the time (h) after incubation. Furthermore,
the time for the maximum rate of digestion (tRM) and thereby
gas production was calculated by the following formula:
1
tRM = B × (C − 1) C
where tRM is the time (h) for maximum rate of digestion, B
is the time (h) after incubation at which half of the MGP is
reached, and C is a constant related to the shape of the asso-
ciated gas production curve. The disappearance of DM after
incubation was calculated as:
Disappearance( % ) =
Feed in bottle (g) − feed residue after incubation (g)
× 100
Feed in bottle (g)
Calculations for the mobile bag experiment
The transit time of the mobile bags from stomach to recovery
in the caecum was calculated using the following equation by
Faichney (1975):
n
Transit time = ti × M i
i=1
where ti is the time elapsed from intubation to the midpoint
of the ith time, and ith is the number of bags found at the
ith time as a fraction of the total number of bags found. The
recovery rate of mobile bags found in the caecum during the
10 h collection period was calculated as:
Lindskov Stang et al.
number of bags found in caecum
Recovery rate (% ) = × 100
total number of bags intubated
The DM or nutrient disappearance was calculated as:
Disappearance( % ) =
Feed in bag (g) − feed residue after collection of bag (g)
× 100
Feed in bag (g)
Statistical analysis
All statistical analyses were performed using MIXED proce-
dure in SAS (Version 9.4). Differences in chemical compo-
sition was analyzed in a model comprising the fixed effects
of harvest time (early, medium, and late), grass species (SB,
MF, CF, PR, TF, and TI), and their interaction. The parameters
derived from the in vitro measurements (dDM, pH, A, B, C,
and tRM) were analyzed in a model comprising the fixed effects
of harvest time (early and late), grass species (SB, MF, CF, PR,
TF, and TI), and their interaction. Dry matter disappearance
from individual mobile bags were plotted against time, and
the slope and intercept from the linear regressions were ana-
lyzed in a model comprising the fixed effects of harvest time
(early and late), grass species (CF and PR), and pore size (15
and 36 µm), no interactions were tested. The disappearance of
nutrients from the mobile bags were analyzed in a model com-
prising the fixed effects of harvest time (early and late), grass
species (CF and PR), pore size (15 and 36 µm), and time (1-3,
4-6, and 7-10 h), no interactions were tested. Finally, the loss
of nutrients from the mobile bags when washed was analyzed
in a model comprising the fixed effects of harvest time (early
and late), grass species (CF and PR), and pore size (15 and 36
µm), no interactions were tested. Horse was not included in
any of the models as inoculum used in the in vitro experiment
was pooled from different horses, and residue from mobile
bags were pooled by each grass species × harvest time × pore
size combination in addition to the three time-intervals where
bags were found in the caecum (1–3, 4−6, and 7–10 h after
bags were administered into the stomach). Results are pre-
sented as least square means with standard error of the mean
(SEM) as a measure of variance. Effects were considered sta-
tistically different if P < 0.05.
Results
Experimental feeds
The chemical composition was affected by grass species and
harvest time as shown in Figure 1. Interactions between
grass species and harvest time (P < 0.001) were present for
all nutrient fractions analyzed (CP, NDF, total WSC, glu-
cose, fructose, sucrose, and fructans). Crude protein content
decreased in all grasses with increasing harvest time (Figure
1). Content of NDF increased with increasing harvesting time
for all species except SB where the NDF content decreased
and TI where the NDF content was similar between harvests
(Figure 1). Total WSC increased for SB and TI, decreased
for CF or did not change with increasing harvest time for
MF, PR, and TF. In general, higher concentrations of WSC
were measured for PR than for the other grass species but
late harvested SB had the same WSC concentrations as PR.
5
The lowest concentration of WSC across all harvests were
present in MF and CF. Glucose concentration decreased in
all grass species from early to late harvest time, and this
was also the case for fructose in CF and PR. Furthermore,
PR had higher levels of glucose and fructose than the other
grass species. Sucrose levels were higher in SB and PR than
the other grass species and in SB the level increased with
increasing harvest time, while the opposite was measured
for CF. Fructan levels were higher in the late harvest com-
pared to the early and medium harvest in SB, PR, TF, and TI,
while in MF and CF, the fructan levels did stay the same at
all harvest times.
Downloaded from https://academic.oup.com/jas/article/doi/10.1093/jas/skac422/6964640 by guest on 08 December 2023
In vitro experiment
The early and late harvested samples were selected for in
vitro analysis based on the effect of harvest time on the
chemical composition. In vitro DM digestibility parame-
ters are presented in Table 1. The in vitro DM digestibility
after 48 h of incubation was higher (P < 0.001) for early
harvested grass than for late harvested grass. The pH after
48 h of incubation was affected by grass species (P < 0.001)
and harvest time (P < 0.001), where PR had the lowest pH
and CF the highest pH. Grass harvested early had lower pH
after 48 h of incubation compared to late harvested grass.
There was an interaction between grass species and harvest
time for A (P < 0.001), B (P < 0.001), and C (P = 0.012)
parameters as well as the tRM (P < 0.001). In general, the A
parameter was higher for early compared to late harvest.
The A parameter was higher for MF, PR, TF, and TI than SB
and CF in the early harvest, and at the late harvest SB and
TF had the highest A value and CF the lowest. Parameter
B was higher for late harvested SB, MF and CF, similar for
TF, and lower for PR and TI compared to early harvest. In
general, parameter C was highest for the early compared to
the late harvest.
Mobile bag experiment
Two grass species (CF and PR) were selected for the mobile bag
experiment based on their clear differences in chemical com-
position (Figure 1) and in vitro fermentation characteristics
(Figure 2). The average transit time of the mobile bags were
4.7 ± 1.7 h, and the recovery of bags in the caecum was 78%
and in total 97% of the bags were found (Supplementary Table
S1). The DM disappearance from the mobile bags increased
over time at the same rate (slope of lines: 0.608) but the inter-
cept was affected by grass species and harvest time (Table 2;
Figure 3) and the intercept was higher (P < 0.001) for the early
than the late harvest time within each of the two grass spe-
cies. Disappearance of DM, CP, WSC, glucose, fructose, and
fructans from the mobile bags were higher (P < 0.01) for PR
than for CF (Table 3). Late harvest time increased (P = 0.014)
the disappearance of WSC and fructose compared to early
harvest time. Disappearance of NDF, CP, WSC, and fructose
was higher (P < 0.05) in bags with pore size 36 compared to
15 µm (Table 3). Nutrients were lost from the mobile bags by
washing (Supplementary Table S2). Washing loss of DM, CP,
WSC, glucose, fructose, and fructans was higher (P < 0.01) in
PR than in CF. Harvest time affected (P < 0.05) the washing
loss of aNDFom, WSC, and fructose, with the early harvest
having lower loss than late harvest, despite the loss was close
to 0 (Supplementary Table S2). Pore size affected (P < 0.05) the
loss of CP, WSC, and fructose with greater loss from bags with
36 µm compared to 15 µm (Supplementary Table S2).
6 Journal of Animal Science, 2023, Vol. 101

Downloaded from https://academic.oup.com/jas/article/doi/10.1093/jas/skac422/6964640 by guest on 08 December 2023

Figure 1. The effect of grass species (meadow fescue (MF), cocksfoot (CF), perennial ryegrass (PR), smooth bromegrass (SB), tall fescue (TF), and
timothy (TI)) harvested at three different plant maturities (early, middle, and late first cut) on nutrient content. Values are presented as means ± SEM.
Letters indicate statistical difference between harvest time (early, medium, and late harvest) within grass species.

Discussion It was demonstrated that harvesting time affected chemical

composition of grasses and their digestibility in different
This study aimed to quantify carbohydrate composition and
ways, but also that different grass species harvested at the
digestion of various grass species harvested at different times.
Lindskov Stang et al. 7

Table 1. The dry matter disappearance (dDM) and pH after 48 h of in vitro incubation and the model parameters A (the asymptotic gas production, mL/g
DM), B (the time (h) after incubation at which half of A is reached), C (constant related to the shape of the curve), and tRM (time for maximum rate of
digestion) of six different grass species (meadow fescue (MF), cocksfoot (CF), perennial ryegrass (PR), smooth bromegrass (SB), tall fescue (TF) and
timothy (TI)) harvested at two different harvesting times (early and late)

Grass species Harvest P-values

SB MF CF PR TF TI Early SEM Late SEM G H G×H

dDM 64.4 63.8 62.3 62.2 60.8 59.8 70.1a 0.75 54.4b 0.70 0.091 <0.001 0.088
pH 6.59bc 6.60b 6.65a 6.56c 6.61b 6.59bc 6.57b 0.01 6.63a 0.01 <0.001 <0.001 NS
A Early 164.5 b
185.5 a
163.1 b
181.8 a
178.2 a
173.0 a
<0.001 <0.001 <0.001
Late 151.8a 140.2bc 112.3d 133.5c 142.0ab 136.9c

Downloaded from https://academic.oup.com/jas/article/doi/10.1093/jas/skac422/6964640 by guest on 08 December 2023

SEM 3.8 10.2 11.5 11.0 8.7 9.1
B Early 9.7ab,x 9.0b,x 8.4b,x 8.3b,x 10.5a,x 9.6ab,x <0.001 <0.001 <0.001
Late 10.2 b,y
12.1 a.y
12.0 a,y
7.6 c,y
10.5 b,x
9.4b,y
SEM 0.34 0.71 0.89 0.21 0.43 0.27
C Early 1.80a,x 1.72ab,x 1.95a,x 1.62b,x 1.38c,x 1.79a,x <0.001 <0.001 0.012
Late 1.40b,y 1.44b,y 1.88a,y 1.45b,y 1.24c,y 1.35bc,y
SEM 0.10 0.07 0.06 0.04 0.04 0.11
tRM Early 8.60a,x 7.38b,x 8.08a,x 6.17c,x 5.11d,x 8.43a,x <0.001 <0.001 <0.001
Late 5.19c,y 6.82b,y 11.02a,y 4.36d,y 3.27e,y 4.25d,y
SEM 0.87 0.24 0.66 0.42 0.44 1.04

Values are presented as means ± SEM. G, grass species; H, harvest time; G × H, interaction effects of grass species × harvest time.
a,b,c,d
Values within a row are different if superscript differs (P < 0.05).
x,y
Values within a column are different if superscript differs (P < 0.05).

same date had different digestibility. A later harvest time of
the different grass species clearly reduced the digestibility as
shown using the IVGPT. Furthermore, the precaecal digest-
ibility of the WSC fraction of PR and CF was high when
quantified using the MBT. The methodological possibilities
and limitations as well as practical application of the results
are discussed in the following.
Experimental feeds
The CP content decreased in all grass species and the aND-
Fom content increased in MF, CF, PR, and TF with later har-
vesting, which is in accordance with results from other studies
performed in comparable climates (Ragnarsson and Lind-
berg, 2008, 2010; Müller, 2012). In the late harvested SB, the
grass had lodged, making it difficult to harvest the lower fiber
rich parts of the grass and this may explain the decrease in
aNDFom from early to late harvest time (Griggs et al., 2007).
The content of WSC, especially glucose and sucrose, in grass
is of relevance for horses with insulin dysregulation (Frank et
al., 2010; Lindåse et al., 2016), pituitary pars intermedia dys-
function (McGowan, 2008), or polysaccharide storage myop-
athy (Firshman et al., 2003; Borgia et al., 2011). Furthermore,
experimentally induced hindgut acidosis using large amounts
of fructans has resulted in development of laminitis (Van Eps
and Pollitt, 2006), although the fructan used in that study was
not of grass origin. In general, WSC was higher for PR than
the other grass species, and fructans increased in SB, PR, TF
and TI at the late harvest but not in MF and CF. In a more
southern location (Utah, USA), Jensen et al. (2014) found PR
to have high WSC, but they also found an increase of fruc-
tans in CF at later harvests. Differences between the studies
may be due to differences in e.g., temperature, photoperiod,
N availability, and cultivar. Variations in content of fructans
have been found between cultivars of CF adapted to different
climates (Sanada et al., 2007), and it would be relevant to test
different cultivars of the different grass species. In contrast to
PR, CF, and MF might therefore be suitable for horses prone
to the abovementioned health conditions because of the rel-
atively low content of WSC, and it should be investigated
further. In all grass species except PR, an early harvest was
also a possibility to reduce levels of WSC and fructans. This
will however also result in a higher overall digestibility of the
grass and a high energy value, which is undesired for horses
with comparably low energy requirements as it reduces the
amount of forage needed to cover energy requirements which
in turn decrease forage intake time (Müller, 2011). Further-
more, it has been found that horses may find CF less palatable
compared to other grass species when grazing (Allen et al.,
2013) and this could potentially also be used to reduce feed
intake in horses.
The in vitro experiment
The in vitro measurements reflect the total tract digestibil-
ity, and in combination with the chemical composition of the
grass samples, the in vitro experiment was used for selection
of grass samples to be used for the mobile bag experiment.
The DM disappearance in the in vitro experiment decreased
from early to late harvest (70.1% vs. 54.4%) which is in
accordance with results from in vivo studies where digestibil-
ity of grass forages in horses decreased with increasing plant
maturity at harvest (Ragnarsson and Lindberg, 2008, 2010;
Müller, 2012). No difference in DM disappearance was found
between grass species within harvest time, but the MGP indi-
cated differences between the grass species. Early and late
harvested CF had the lowest MGP indicating a generally
low digestibility of CF. The MGP differed more between late
than early harvested grass samples, and SB had the highest
MGP at the late harvest probably due to a high content of
WSC. Differences in nutrient content results in different in
vitro fermentations and model parameters (A, B, C, and tRM)
8 Journal of Animal Science, 2023, Vol. 101

Downloaded from https://academic.oup.com/jas/article/doi/10.1093/jas/skac422/6964640 by guest on 08 December 2023

Figure 2. Measured (dotted lines) and modeled (solid line) in vitro gas production (ml/g organic matter (OM)) when incubating three replicates of early
and late cut grasses (meadow fescue (MF), cocksfoot (CF), perennial ryegrass (PR), smooth bromegrass (SB), tall fescue (TF) and timothy (TI)) for 48 h
using horse caecal fluid as inoculum. Early cut grasses had higher in vitro gas production than late cut grasses.

the hindgut. Hence, a predigestion step before the in vitro fer-

Table 2. The linear regression analysis of the relation between grass
mentation would be required. This has not been standardized
species (cocksfoot (CF) or perennial ryegrass (PR)) and harvest time
(early or late) on dry matter disappearance (dDM) from mobile bags over for horses, but it should be considered in future studies. A pre-
time (X, hours). digestion step would also be useful for quantifying the in vitro
prececal digestion of feedstuffs like the method suggested by
Grass species Harvest Equation Tilley and Terry (1963). Still, MGP and DM disappearance
give an indication of the potential total tract digestibility of
PR Early dDM = 0.608 X + 38.01a the grass species. The pH values measured after 48 h of in
Late dDM = 0.608 X + 32.51b vitro incubation cannot be extrapolated directly to in vivo
CF Early dDM = 0.608 X + 35.29ab hindgut pH values as the buffer solution will buffer the in
Late dDM = 0.608 X + 21.37c vitro system. However, a higher DM disappearance with a
potentially higher production of SCFA was reflected in a
Values within a column are different if superscript differs (P < 0.05).
a,b,c,d
lower pH. The traditional digestibility studies with total col-
lection of feces gives a more accurate estimate of the apparent
total tract digestibility, but it does not distinguish where in
differed between grasses. However, not all these parameters the gastrointestinal tract digestion takes place, similar to the
can be transferred directly to the gastrointestinal physiology IVGPT. In this study, the in vitro measurements were used
of the horse. Here the feed undergoes enzymatic digestion in successfully to screen a large number of grass samples, as not
the stomach and small intestine before residual nutrients not all could be tested in vivo using the MBT. Based on the chem-
absorbed in the small intestine are potentially fermented in ical composition and in vitro measurements, CF and PR were
Lindskov Stang et al. 9

Table 3. The effect of grass species (cocksfoot (CF) or perennial ryegrass (PR)), harvest time (early or late), pore size of mobile bags (15 or 36 µm), and
time of recovery of mobile bags (1–3, 4–6 or 7–10 h) after administration on nutrient disappearance (%) from mobile bags

Grass species Harvest time Pore size Time P-values

CF SEM PR SEM Early SEM Late SEM 15 SEM 36 SEM 1-3 SEM 4-6 SEM 7-10 SEM G H P T

DM 31.3 2.2 38.3 1.0 39.8 0.7 29.8 1.8 33.9 2.0 35.7 2.0 33.0 2.5 35.0 2.4 36.5 2.5 <0.001 <0.001 0.098 0.036
aND- 7.0 1.0 6.4 0.7 8.8 0.8 4.5 0.4 5.7 0.8 7.7 0.9 5.1 0.7 6.5 1.1 8.4 1.1 0.29 <0.001 <0.001 <0.001
Fom
CP 71.6 1.3 77.0 1.7 78.1 1.3 70.4 1.3 73.2 1.7 75.2 1.7 70.0 1.9 75.8 2.1 77.0 1.6 <0.001 <0.001 <0.001 <0.001
WSC 98.5 0.3 99.2 0.2 99.1 0.2 98.6 0.3 98.7 0.4 99.1 0.2 97.8b 0.4 99.2a 0.1 99.7a 0.2 <0.01 0.028 0.056 <0.001
Glu- 99.0 0.4 99.0 0.2 98.7 0.3 99.4 0.2 98.9 0.3 99.2 0.3 98.6 0.6 99.0 0.2 99.5 0.3 0.98 0.14 0.41 0.24

Downloaded from https://academic.oup.com/jas/article/doi/10.1093/jas/skac422/6964640 by guest on 08 December 2023

cose
Fruc- 99.1 0.2 99.4 0.2 99.6 0.1 98.9 0.2 99.2 0.2 99.2 0.2 98.7 0.2 99.4 0.1 99.6 0.2 0.062 <0.001 0.96 <0.001
tose
Sucrose 93.2 1.8 99.2 0.4 98.1 0.8 94.2 1.9 95.5 1.8 96.9 1.4 93.6 2.7 97.8 0.8 97.2 1.7 <0.01 0.025 0.37 0.098
Fructan 99.4 0.4 98.7 0.3 99.0 0.4 99.1 0.4 99.1 0.3 99.0 0.4 98.2 0.6 99.5 0.2 99.5 0.2 0.12 0.76 0.76 0.027

Values are presented as means ± SEM. G, grass species; H, harvest time; P, pore size; T, time of recovery of mobile bags.
DM, dry matter; CP, crude protein; WSC, water-soluble carbohydrates; aNDFom, neutral detergent.
Fiber assayed with heat-stable amylase and expressed without residual ash.

Figure 3. Linear regression of the dry matter disappearance from mobile bags with a pore size of 15 or 36 µm containing early and late harvested
perennial ryegrass (PR) and cocksfoot (CF).

selected as highly contrasting grasses with distinct differences
in WSC content and DM disappearance.
The mobile bag experiment
The disappearance of DM, CP, and aNDFom decreased from
early to late harvest in accordance with results from other
studies where digestibility decreased with increasing plant
maturity (Ragnarsson and Lindberg, 2008, 2010; Müller,
2012; Särkijärvi et al., 2012). The effect of harvest time on
dDM was more pronounced for CF than PR, which might
be explained by a faster morphological development resulting
in a greater increase in aNDFom across harvest times for CF
compared to PR. It was found that the precaecal disappear-
ance of aNDFom was low, as the precaecal fermentation has
been described as limited (Moore-Colyer et al., 2002; Varloud
et al., 2004; Brøkner et al., 2012). The precaecal disappear-
ances of glucose, sucrose and fructose were all above 93%
which was expected, but there are no published studies (to
the knowledge of the authors) where MBT has been used to
study the digestion of WSC in horses previously. The high pre-
caecal disappearance of fructans (>98%) was however not
expected. Nutrients disappearing from the mobile bags might
not be digested and absorbed in the small intestine, but these
results indicate that grass fructans are easily available for deg-
radation precaecally, and it is probable that microbes in the
stomach and/or small intestine ferment fructans in these com-
partments. Additionally, a relatively large proportion (more
than ~70%) of the fructans disappeared from the mobile
bags by washing them (Supplementary Table S2), confirm-
ing that fructans are easily available for degradation precae-
cally. A preliminary study has shown that a part of fructans
is digested in the stomach of the horse (Coenen et al. 2006),
probably through acid hydrolysis from the acid environment
in the lower part of the stomach. Results from in vitro studies
(Ince et al., 2014; Strauch et al., 2017) have also indicated
that fructans might be degraded precaecally by acidic hydro-
lysis and/or fermentation. Knowledge on which of the WSC
fractions that gives the lowest responses on blood glucose and
insulin levels would be relevant to investigate in future studies
in relation to horses with e.g., insulin dysregulation.
Mobile bags were recovered at different time points after
administration in the current study. This is of importance
as a longer time within the gastrointestinal tract can poten-
tially increase the disappearance of nutrients (Hymøller et al.,
2012; Thorringer et al., 2020). Mobile bags were therefore
pooled according to time of recovery and not horses in this
study, as suggested by Thorringer et al. (2020). A linear rela-
tionship between disappearance of DM and time to recovery
of bags was found, however, the slope of the curves did not
differ, only the intercept. Early harvested grass had a greater
10
i­ntercept than late harvested grass due to higher content of
soluble nutrients such as WSC and CP. This effect of time
should be addressed in studies using the MBT as recovery
time of bags might vary (could be affected by e.g., diet compo-
sition and individual horse). Another methodological aspect
of the mobile bag technique is the pore size of the bags. The
technique is not standardized, and different pore sizes and
FSA (mg feed/cm2) have been used previously (Moore-Colyer
et al., 2002; Thorringer et al., 2020). The Nordic feed evalua-
tion system for ruminants (Åkerlind et al., 2011) recommends
a pore size of 15 μm and a FSA of 5–7 and 10–15 mg feed/
cm2 for roughages and concentrates, respectively. Thorringer
and Jensen (2021) recommends not to use a FSA of more than
20 mg/cm2, to still have enough residue for analyses and not
compromising the disappearance of nutrients from the bags.
In this study, pore size influenced some nutrients, and it is
recommended to use a pore size of 15 μm rather than 36 μm
in future studies investigating the precaecal digestion of grass
and other fibrous feedstuffs.
Conclusion
This study demonstrated that both grass species and harvest-
ing time affect the chemical composition of the grasses. In gen-
eral, the CP decreased and aNDFom increased as the grasses
matured. The content of WSC increased in SB and TI, but
decreased in CF, and fructans increased in SB, TI, PR, and TF
as they matured. The in vitro technique used showed clearer
differences between harvesting time than between grass spe-
cies. This technique can be used to screen a large number of
samples within a short time. However, the in vitro fermenta-
tion does not reflect the gastrointestinal physiology of horses
and a predigestion step should be considered in future stud-
ies. The mobile bag technique showed a high precaecal disap-
pearance of the WSC fraction in grasses, including fructans.
However, nutrients disappearing from the mobile bags are
not necessarily digested precaecally, but the results show that
fructans might be accessible for microbial fermentation and/
or acid hydrolysis before reaching the hindgut. Harvesting
time had larger effect than grass species on digestibility, but
different grass species also influenced digestibility and com-
position. More specifically, CF consistently had lower and PR
higher digestibility and WSC content. Further studies on grass
species suitable for different horse categories should include
the differences between grass species and focus on how plant
maturity at harvest can be used to achieve the desired nutrient
content and digestibility in the forage.
Supplementary Data
Supplementary data are available at Journal of Animal Science
online.
Supplementary Figure S1. Weather data showing daily av-
erage, minimum (Min.) and maximum (Max.) temperature
(temp.) in °C as well as daily rain in mm during the exper-
iment. Vertical lines indicate days for harvesting early, me-
dium, and late first cut samples. Historical weather data is
obtained from an online weather service (www.yr.no).
Acknowledgments
This work was financially supported by The Swedish-
Norwegian Foundation for Equine Research (grant number
Journal of Animal Science, 2023, Vol. 101
H-17-47-287). We gratefully acknowledge Mette Henne and
Øyvind Jørgensen for technical assistance during the animal
trials and production of experimental feeds, respectively.
Conflict of Interest Statement
The authors declare that there are no conflicts of interest.
Literature Cited
Åkerlind, M., M. Weisbjerg, T. Eriksson, R. Tøgersen, P. Udén, B. L.
Olafsson, O. M. Harstad, and H. Volden. 2011. Feed analyses and
Downloaded from https://academic.oup.com/jas/article/doi/10.1093/jas/skac422/6964640 by guest on 08 December 2023
digestion methods. In: Volden, H., editor. NorFor: the Nordic feed
evaluation system. Wageningen, The Netherlands: Wageningen Ac-
ademic Publishers; p. 41–54.
Allen, E., C. Sheaffer, and K. Martinson. 2013. Forage nutritive value
and preference of cool-season grasses under horse grazing. Agron.
J. 105:679–684. doi:10.2134/agronj2012.0300
ANKOM. 2022. ANKOM RF Gas Production System Operator’s Man-
ual <900 [accessed October 31, 2022]. Available from https://www.
ankom.com/product-catalog/ankom-rf-gas-production-system.
Borgia, L., S. Valberg, M. McCue, K. Watts, and J. Pagan. 2011.
Glycaemic and insulinaemic responses to feeding hay with
different non-structural carbohydrate content in control and
polysaccharide storage myopathy-affected horses. J. Anim.
Physiol. Anim. Nutr. (Berl). 95:798–807. doi:10.1111/j.1439-
0396.2010.01116.x
Brøkner, C., D. Austbø, J. A. Næsset, K. E. B. Knudsen, and A.-H.
Tauson. 2012. Equine pre-caecal and total tract digestibility of
individual carbohydrate fractions and their effect on caecal pH
response. Arch. Anim. Nutr. 66:490–506. doi:10.1080/174503
9x.2012.740311
Coenen, M., A. Mösseler, and I. Vervuert. 2006. Fermentative gases in
breath indicate that inulin and starch start to be degraded by mi-
crobial fermentation in the stomach and small intestine of the horse
in contrast to pectin and cellulose. J. Nutr. 136:2108S–2110S.
doi:10.1093/jn/136.7.2108S
Dyer, J., E. F.-C. Merediz, K. S. H. Salmon, C. J. Proudman, G. B. Ed-
wards, and S. P. Shirazi-Beechey. 2002. Molecular characterisation
of carbohydrate digestion and absorption in equine small intestine.
Equine Vet. J. 34:349–358. doi:10.2746/042516402776249209
Faichney, G. J. 1975. The use of markers to partition digestion within
the gastro-intestinal tract of ruminants. Digestion and metabolism
in the ruminant 1974. In: Proceedings of the 4th International Sym-
posium on Ruminant Physiology. Armidale, NSW, Australia: Univ.
of New England Publishing Unit; p. 277–291.
Firshman, A. M., S. J. Valberg, J. B. Bender, and C. J. Finno. 2003. Epi-
demiologic characteristics and management of polysaccharide stor-
age myopathy in Quarter Horses. Am. J. Vet. Res. 64:1319–1327.
doi:10.2460/ajvr.2003.64.1319
Frank, N., R. J. Geor, S. R. Bailey, A. E. Durham, and P. J. Johnson.
2010. Equine metabolic syndrome. J. Vet. Intern. Med. 24:467–
475. doi:10.1111/j.1939-1676.2010.0503.x
Goering, H. K., and P. J. Van Soest. 1970. Forage fiber analyses (Appa-
ratus reagents, procedures, and some applications). Agric. Hand-
book No. 379. Washington D: ARS USDA; p. 20.
Griggs, T. C., J. W. MacAdam, H. F. Mayland, and J. C. Burns. 2007.
Temporal and vertical distribution of nonstructural carbohydrate,
fiber, protein, and digestibility levels in orchardgrass swards. Agron.
J. 99:755–763. doi:10.2134/agronj2006.0036
Groot, J. C. J., J. W. Cone, B. A. Williams, F. M. A. Debersaques, and E.
A. Lantinga. 1996. Multiphasic analysis of gas production kinetics
for in vitro fermentation of ruminant feeds. Anim. Feed Sci. Tech-
nol. 64:77–89. doi:10.1016/S0377-8401(96)01012-7
Harris, P. A., A. D. Ellis, M. J. Fradinho, A. Jansson, V. Julliand, N.
Luthersson, A. S. Santos, and I. Vervuert. 2017. Review: feeding
conserved forage to horses: recent advances and recommendations.
Animal. 11:958–967. doi:10.1017/S1751731116002469
Lindskov Stang et al.
Humphreys, M. W., R. S. Yadav, A. J. Cairns, L. B. Turner, J. Humphreys,
and L. Skøt. 2006. A changing climate for grassland research. New
Phytol. 169:9–26. doi:10.1111/j.1469-8137.2005.01549.x
Hymøller, L., M. S. d. Dickow, C. Brøkner, D. Austbø, and S. K.
­Jensen. 2012. Cereal starch, protein, and fatty acid pre-caecal dis-
appearance is affected by both feed technological treatment and
­efficiency of the chewing action in horses. Livest. Sci. 150:159–169.
doi:10.1016/j.livsci.2012.08.016
Ince, J. C., A. C. Longland, M. J. S. MooreColyer, and P. A. Harris.
2014. In vitro degradation of grass fructan by equid gastrointesti-
nal digesta. Grass Forage Sci. 69:514–523. doi:10.1111/gfs.12061
Janis, C. 1976. The evolutionary strategy of the equidae and the or-
igins of rumen and cecal digestion. Evolution. 30:757–774.
doi:10.2307/2407816
Jensen, K. B., P. Harrison, N. J. Chatterton, B. S. Bushman, and J. E.
Creech. 2014. Seasonal trends in non-structural carbohydrates
in cool- and warm-season grasses. Crop Sci. 54:2328–2340.
doi:10.2135/cropsci2013.07.0465
King, C., J. McEniry, M. Richardson, and P. O’Kiely. 2012. Yield and
chemical composition of five common grassland species in response
to nitrogen fertiliser application and phenological growth stage.
Acta Agric. Scand. B Soil Plant Sci. 62:644–658. doi:10.1080/09
064710.2012.687055
Lindåse, S., C. Müller, K. Nostell, and J. Bröjer. 2018. Evaluation of glu-
cose and insulin response to haylage diets with different content of
nonstructural carbohydrates in 2 breeds of horses. Domest. Anim.
Endocrinol. 64:49–58. doi:10.1016/j.domaniend.2018.03.006
Lindåse, S. S., K. E. Nostell, C. E. Müller, M. Jensen-Waern, and J. T.
Bröjer. 2016. Effects of diet-induced weight gain and turnout to
pasture on insulin sensitivity in moderately insulin resistant horses.
Am. J. Vet. Res. 77:300–309. doi:10.2460/ajvr.77.3.300
Longland, C. A., and B. M. Byrd. 2006. Pasture nonstructural car-
bohydrates and equine laminitis. J. Nutr. 136:2099S–2102S.
doi:10.1093/jn/136.7.2099S
Lowman, R. S., M. K. Theodorou, J. J. Hyslop, M. S. Dhanoa, and D. Cud-
deford. 1999. Evulation of an in vitro batch culture technique for esti-
mating the in vivo digestibility and digestible energy content of equine
feeds using equine faeces as the source of microbial inoculum. Anim.
Feed Sci. Tech. 80:11–27. doi:10.1016/S0377-8401(99)00039-5
McGowan, C. 2008. The role of insulin in endocrinopathic laminitis. J.
Equine Vet. Sci. 28:603–607. doi:10.1016/j.jevs.2008.08.004
Merediz, E. F., J. Dyer, K. S. H. Salmon, and S. P. Shirazi-Beechey. 2004. Mo-
lecular characterisation of fructose transport in equine small intes-
tine. Equine Vet. J. 36:532–538. doi:10.2746/0425164044877378
Moore-Colyer, M. J. S., J. J. Hyslop, A. C. Longland, and D. Cudd-
eford. 2002. The mobile bag technique as a method for determin-
ing the degradation of four botanically diverse fibrous feedstuffs in
the small intestine and total digestive tract of ponies. Br. J. Nutr.
88:729–740. doi:10.1079/bjn2002734
Müller, C. E. 2011. Equine ingestion of haylage harvested at differ-
ent plant maturity stages. Appl. Anim. Behav. Sci. 134:144–151.
doi:10.1016/j.applanim.2011.08.005
Müller, C. E. 2012. Equine digestion of diets based on haylage harvest-
ed at different plant maturities. Anim. Feed Sci. Tech. 177:65–74.
doi:10.1016/j.anifeedsci.2012.06.002
Philippeau, C., M. Varloud, and V. Julliand. 2014. Mobile bag starch
prececal disappearance and postprandial glycemic response of four
11
forms of barley in horses. J. Anim. Sci. 92:2087–2093. doi:10.2527/
jas.2013-6850
Pollock, C. J., and A. J. Cairns. 1991. Fructan metabolism in grasses
and cereals. Annu. Rev. Plant Physiol. Plant Mol. Biol. 42:77–101.
doi:10.1146/annurev.pp.42.060191.000453
Ragnarsson, S., and J. E. Lindberg. 2008. Nutritional value of timothy
haylage in Icelandic horses. Livest. Sci. 113:202–208. doi:10.1016/j.
livsci.2007.03.010
Ragnarsson, S., and J. E. Lindberg. 2010. Nutritional value of
mixed grass haylage in Icelandic horses. Livest. Sci. 131:83–87.
doi:10.1016/j.livsci.2010.03.003
Ringmark, S., L. Roepstorff, B. Essèn-Gustavsson, T. Revold, A. Lind-
holm, U. Hedenström, M. Rundgren, G. Ögren, and A. Jansson.
Downloaded from https://academic.oup.com/jas/article/doi/10.1093/jas/skac422/6964640 by guest on 08 December 2023
2013. Growth, training response and health in Standardbred
yearlings fed a forage-only diet. Animal. 7:746–753. doi:10.1017/
s1751731112002261
Sanada, Y., T. Tomoyuki, and T. Yamada. 2007. Ecotypic variation of
water-soluble carbohydrate concentration and winter hardiness
in cocksfoot (Dactylis glomerata L.). Euphytica. 153:267–280.
doi:10.1007/s10681-006-9262-9
Särkijärvi, S., R. Sormunen-Cristian, T. Heikkilä, M. Rinne, and M.
Saastamoinen. 2012. Effect of grass species and cutting time on
in vivo digestibility of silage by horses and sheep. Livest. Sci.
144:230–239. doi:10.1016/j.livsci.2011.11.019
Shewmaker, G. E., H. F. Mayland, C. A. Roberts, P. A. Harrison, N. J.
Chatterton, and D. A. Sleper. 2006. Daily carbohydrate accumu-
lation in eight tall fescue cultivars. Grass Forage Sci. 61:413–421.
doi:10.1111/j.1365-2494.2006.00550.x
Strauch, S., B. Wichert, J. M. Greef, D. Hillegeist, A. Zeyner, and A. Lie-
segang. 2017. Evaluation of an in vitro system to simulate equine
foregut digestion and the influence of acidity on protein and fruc-
tan degradation in the horseʹs stomach. J. Anim. Physio. Anim.
Nutr. 101:51–58. doi:10.1111/jpn.12635
Thorringer, N. W., and R. B. Jensen. 2021. Methodical considerations
when estimating nutrient digestibility in horses using the mobile bag
technique. Animal. 15:100050. doi:10.1016/j.animal.2020.100050
Thorringer, N. W., M. R. Weisberg, and R. B. Jensen. 2020. The effects
of processing barley and maize on metabolic and digestive respons-
es in horses. J. Anim. Sci. 98:skaa353. doi:10.1093/jas/skaa353
Thorringer, N. W., M. R. Weisberg, and R. B. Jensen. 2022. Mobile
bag technique for estimation of nutrient digestibility when hay is
supplemented with alternative fibrous feedstuffs in horses. Anim.
Feed Sci. Tech. 283:115168. doi:10.1016/j.anifeedsci.2021.115168
Tilley, J. M. A., and R. A. Terry. 1963. A two-stage technique for the
in vitro digestion of forage crops. Grass Forage Sci. 18:104–111.
doi:10.1111/j.1365-2494.1963.tb00335.x
Udén, P. 2006. In vitro studies on microbial efficiency from two cuts
of ryegrass (Lolium perenne cv. Aberdart) with different propor-
tions of sugars and proteins. Anim. Feed Sci. Tech. 126:145–156.
doi:10.1016/j.anifeedsci.2005.06.008
Van Eps, A. W., and C. C. Pollitt. 2006. Equine laminitis induced with
oligofructose. Equine Vet. J. 38:203–208. doi:10.2746/0425164067
76866327
Varloud, M., A. de Fombelle, A.-G. Goachet, C. Drogoul, and V. Jul-
liand. 2004. Partial and total apparent digestibility of dietary car-
bohydrates in horses as affected by the diet. Anim. Sci. 79:61–72.
doi:10.1017/S1357729800054539